An international team of researchers led by NIMS managed to produce a neuromorphic network composed of several metallic nanowires.
Utilizing this network, the scientists could generate electrical signs similar to those linker with higher-order brain functions found only in humans, such as memorization, learning, forgetting, and so on. The team then refined the processes that generated these electrical characteristics.
The progress of artificial intelligence (AI) methods has been quickly advancing over the last year and has begun affecting our lives in numerous ways. Even though AI processes data in a similar way to the human brain, the mechanisms by which human brain function are still mostly unknown. Basic brain elements, such as neurons and the synapses (junctions between them), have been closely analyzed.
Even so, several questions regarding the brain as a collective whole are still unanswered. The joint research team recently developed an intricate network similar to the human brain by incorporating numerous silver (Ag) nanowires covered in a polymer (PVP) insulating layer about one nanometer in thickness.
A connection between two nanowires creates a variable resistive factor, for instance, a synaptic element that acts as a neuronal junction. This nanowire network, which includes a wide array of complexly interacting junction elements, creates a ‘neuromorphic network.’
A Memory Device to Replace Computers
When the team applied a voltage to this network, it seemed to ‘struggle’ to find the best current pathways, most precisely, the most electrically effective pathways. The scientists calculated the processes of current pathway development, retention, and deactivation while the electric current was circulating through the network and discovered that these procedures always oscillate as they progress, just like the human brain’s memorization, learning and forgetting processes.
The research team is now creating a memory device similar to the human brain, using the neuromorphic network material. They also intend to develop the memory device to function utilizing fundamentally different standards than those used in today’s computers.
For instance, while computers are now designed to use as much time and electricity as needed in the search for absolutely best solutions, the new device is intended to make a rapid decision within specific restrictions, although the solution given may not be the best.